Documentation/Nightly/Extensions/3D Model Segmentation - Revision history

Abeers: Grant updates

2017-07-20T22:09:25Z

Grant updates

Abeers: Reformatting

2017-07-19T18:43:06Z

Reformatting

Abeers: First Draft of Documentation page

2017-07-19T15:16:03Z

First Draft of Documentation page

Abeers: Draft edits.

2017-07-11T04:11:39Z

Draft edits.

Abeers: Draft edits

2017-07-10T22:36:28Z

Draft edits

Abeers: Brief doc edits

2017-07-10T15:27:25Z

Brief doc edits

Abeers: Creating a template page for the new 3D Model Segmentation extension.

2017-07-10T15:24:20Z

Creating a template page for the new 3D Model Segmentation extension.

New page

<noinclude>{{documentation/versioncheck}}</noinclude>

{{documentation/{{documentation/version}}/module-header}}



{{documentation/{{documentation/version}}/module-section|Introduction and Acknowledgements}}
{{documentation/{{documentation/version}}/module-introduction-start|{{documentation/modulename}}}}
{{documentation/{{documentation/version}}/module-introduction-row}}

{|
|
[[Image:BLANK.png]]
|
Extension: [[Documentation/{{documentation/version}}/Extensions/3D_Model_Segmentation|3D_Model_Segmentation]] 
Acknowledgments:
This work supported in part the National Institutes of Health, National Cancer Institute through the following grants:
* Quantitative MRI [...] 
Authors: Andrew Beers (Massachusetts General Hospital) 
Contact: Andrew Beers, <email>abeers@mgh.harvard.edu</email> 

License: [http://www.slicer.org/pages/LicenseText Slicer License]
|}

{{documentation/{{documentation/version}}/module-introduction-logo-gallery
|{{collaborator|logo|qiicr}}|{{collaborator|longname|qiicr}}
}}

{{documentation/{{documentation/version}}/module-introduction-end}}


{{documentation/{{documentation/version}}/module-section|Module Description}}

{|
|Module description here.
 
[[File:DistanceMapBasedRegistration_example1.png|thumb|400px| Example visualization of MRI/TRUS registration result obtained using this module. Yellow outline is the contour of the manual segmentation of the prostate gland in TRUS, over the registered MRI image of the same prostate.]]
|}


{{documentation/{{documentation/version}}/module-section|Use Cases}}
* Tumor Segmentation


{{documentation/{{documentation/version}}/module-section|Tutorials}}


{{documentation/{{documentation/version}}/module-section|Panels and their use}}
{|
|
*'''Fixed image''' (optional): image that should be used as reference in registration
*'''Segmentation of the fixed image'''
*'''Moving image''' (optional): image to be registered
*'''Segmentation of the moving image'''
*'''Registration affine transform''': transform to store the result of affine registration
*'''Registration deformable transform''': transform to store the result of deformable registration
*'''Visualization''': initializes slice viewers to show overlay of the fixed and moving images (before or after registration), outline of the fixed image segmentation and the registration transform grid and visualization of the fixed and moving segmentation label surfaces in 3d view
|[[File:DistanceMapBasedRegistration_GUI.png|thumb|400px| Module control panel.]]
|}




{{documentation/{{documentation/version}}/module-section|Similar Modules}}
Slicer Registration modules and extensions.


{{documentation/{{documentation/version}}/module-section|References}}
[1] Fedorov A, Khallaghi S, Antonio Sánchez C, Lasso A, Fels S, Tuncali K, Neubauer Sugar E, Kapur T, Zhang C, Wells W, Nguyen PL, Abolmaesumi P, Tempany C. (2015) Open-source image registration for MRI–TRUS fusion-guided prostate interventions. Int J CARS: 1–10. Available: http://link.springer.com/article/10.1007/s11548-015-1180-7.

[2] Fedorov A, Nguyen PL, Tuncali K, Tempany C. (2015). Annotated MRI and ultrasound volume images of the prostate. Zenodo. http://doi.org/10.5281/zenodo.16396


{{documentation/{{documentation/version}}/module-section|Information for Developers}}
* Source code: https://github.com/SlicerProstate/SlicerProstate/tree/master/DistanceMapBasedRegistration


{{documentation/{{documentation/version}}/module-footer}}

@@ Line 15: / Line 15: @@
 Acknowledgments:
 This work supported in part the National Institutes of Health, National Cancer Institute through the following grants:
-* Quantitative MRI [...]<br>
+* Quantitative Image Informatics for Cancer Research (QIICR) (U24CA180918)
 Authors: Andrew Beers (Massachusetts General Hospital)<br>
 Contact: Andrew Beers, <email>abeers@mgh.harvard.edu</email><br>
@@ Line 49: / Line 51: @@
 {{documentation/{{documentation/version}}/module-section|Tutorials}}
 https://www.youtube.com/watch?v=_7oZygGp2ds&t=13s
 <!-- ---------------------------- -->

@@ Line 55: / Line 55: @@
+|
 == Step 1. Volume Selection ==
-*'''Volume Selection''': Choose which image you would like to segment. If you only select one image, steps 2 and 3 (Registration, Normalization, and Subtraction) will be skipped.
+{|
-<gallery>
+|<gallery>
 D_Model_Segmentation_Step_1.PNG|Step 1 User Interface
 </gallery>
 == Step 2. Registration (Optional) ==
 *'''Registration Order''': Choose which order you would like to register your images to. The moving image will be resampled into the space of the fixed image using Slicer's BRAINSFit registration algorithm.
 *'''Registration Method''': Choose which registration method you would like to use. "No Registration" allows you to skip this step. "Rigid Registration" will attempt to rotate and translate the moving image into the space of the fixed image. "Affine Registration" will rotate, translate, scale, and skew the input image. Both affine and rigid registration methods will produce a 3D Slicer Linear Transformation object which you can inspect in the "Transforms" module. "Deformable Registration" will perform deformable BSpline registration on the moving image, and create a BSpline Transform object. In general, each method will be more likely to be accurate but also more time-consuming than the previous method.
 *'''Registration Output''': Either create a new volume for the post-registration moving image, or replace the original image. The new volume will have the name "[VolumeName1]_reg_[VolumeName2]".
-<gallery>
+|}
-D_Model_Segmentation_Step_2.PNG|Step 2 User Interface
+== Step 3. Normalization and Subtraction (Optional) ==
 </gallery>
-== Step 3. Normalization and Subtraction (Optional) ==
 *'''Normalization Methods''': Choose the intensity normalization method for your two volumes. This may help to produce better contrast when calculating a subtraction map.
 *'''Normalization Order''': Choose whether to normalize pre- to post-contrast, or vice versa.
 *'''Calculate Subtraction Map''': Subtract Volume 1 from Volume 2 using Slicer's "Subtract Scalar Volumes" module. The output subtraction volume will be the basis for segmentation in future steps. The output volume's name will be displayed as "[VolumeName2]_subtraction".
-<gallery>
+|}
 == Step 4. ROI ==
-*'''ROI Toolbar''': Use the ROI Toolbar to create a smooth, convex segmentation around your region of interest. Use the Model Marker Placement Tool to lay down border points. Once you place more than three border points, a best-guess segmentation will automatically be created from your border points. You can add as many border points as needed, with the segmentation guess automatically adjusting as necessary. You can move existing border points by de-selecting the Model Marker Placement Tool, and clicking and dragging using the basic mouse tool. The model will update accordingly.
+{|
-*'''Model-Based ROI''': Use this drop-down menu to switch between different sets of markup points. Additional points added with the ROI toolbar will update whichever model is selected in this drop-down menu. Pre-existing sets of points can also be loaded into this menu and turned into the template for a segmentation.
+|<gallery>
 D_Model_Segmentation_Step_4A.PNG|Step 4 User Interface
 D_Model_Segmentation_Step_4B.PNG|Step 4 Example ROI from Markups
 </gallery>
 == Step 5. Thresholding ==
-*'''Threshold Range''': Adjust this slider to set a simple intensity threshold on the smooth segmentation created in the previous step. Voxels in the newly-created label named "[VolumeName1]_roi_label_thresholded" with intensities outside of the slider's bounds will be erased. The original segmentation without any thresholding is available under a second label named "[VolumeName1]_roi_label", and a version of the original volume cropped to that label is available under the name "[VolumeName1]_roi_cropped".
+{|
-<gallery>
+|<gallery>
 D_Model_Segmentation_Step_5A.PNG|Step 5 User Interface
 D_Model_Segmentation_Step_5B.PNG|Step 5 Example Thresholded ROI
 </gallery>
 == Step 6. Review ==
 *'''Editor Tools''': Use these tools from Slicer's Editor module to clean up your segmentation. You can also switch to the Editor or Segmentations modules for more label editing options.
 *'''Restart''': Remove segmentation objects you no longer need, and go back to Step 1. While you can go back to previous steps via the "Back" buttons, it is recommended that you use the Restart button in this step when segmenting an entirely new volume. This will reset all of the steps back to their original state.
-<gallery>
+|}
 <!--
 {{documentation/{{documentation/version}}/module-parametersdescription}}

@@ Line 10: / Line 10: @@
 {|
+[[File:3D Model Segmentation.png]]
+|
 Extension: [[Documentation/{{documentation/version}}/Extensions/3D_Model_Segmentation|3D_Model_Segmentation]]<br>
@@ Line 33: / Line 32: @@
 {|
-|Module description here.
+The 3D Model Segmentation module allows users to quickly create smooth, 3D regions of interest with only a few clicks, and then provides a variety of tools to threshold and refine those regions of interest. The module is broken up into several steps, which the user can progress through with "Back" and "Next" buttons on the bottom of the module panel. The module can be used on single volumes, or on pre- and post-contrast pairs. In the latter case, utilities for registering, normalizing, and subtracting volumes are provided. Check the tutorials section for some usage examples, and send any bugs, comments, or feature requests to abeers@mgh.harvard.edu.
 <br>
 |}
 <!-- ---------------------------- -->
 {{documentation/{{documentation/version}}/module-section|Use Cases}}
-* Tumor Segmentation
+* Tissue segmentation with and pre- and post-contrast images.
 <!-- ---------------------------- -->
@@ Line 52: / Line 56: @@
 == Step 1. Volume Selection ==
 *'''Volume Selection''': Choose which image you would like to segment. If you only select one image, steps 2 and 3 (Registration, Normalization, and Subtraction) will be skipped.
 == Step 2. Registration (Optional) ==
 *'''Registration Order''': Choose which order you would like to register your images to. The moving image will be resampled into the space of the fixed image using Slicer's BRAINSFit registration algorithm.
-*'''Registration Method''': Choose which registration method you would like to use. "No Registration" allows you to skip this step. "Rigid Registration" will attempt to rotate and translate the moving image into the space of the fixed image. "Affine Registration" will rotate, translate, scale, and skew the input image. Both affine and rigid registration methods will produce a 3D Slicer Linear Transformation object which you can inspect in the "Transforms" module. "Deformable Registration" will perform deformable BSpline registration on the moving image, and correspondingly create a BSpline Transform object. In general, each method will be more accurate but more time-consuming than the previous method.
+*'''Registration Method''': Choose which registration method you would like to use. "No Registration" allows you to skip this step. "Rigid Registration" will attempt to rotate and translate the moving image into the space of the fixed image. "Affine Registration" will rotate, translate, scale, and skew the input image. Both affine and rigid registration methods will produce a 3D Slicer Linear Transformation object which you can inspect in the "Transforms" module. "Deformable Registration" will perform deformable BSpline registration on the moving image, and create a BSpline Transform object. In general, each method will be more likely to be accurate but also more time-consuming than the previous method.
 *'''Registration Output''': Either create a new volume for the post-registration moving image, or replace the original image. The new volume will have the name "[VolumeName1]_reg_[VolumeName2]".
 == Step 3. Normalization and Subtraction (Optional) ==
-*'''Normalization Methods''': Choose the method with which to normalize intensities between your two volumes. This may help to produce better contrast when calculating a subtraction map.
+*'''Normalization Methods''': Choose the intensity normalization method for your two volumes. This may help to produce better contrast when calculating a subtraction map.
 *'''Normalization Order''': Choose whether to normalize pre- to post-contrast, or vice versa.
 == Step 4. ROI ==
+*'''ROI Toolbar''': Use the ROI Toolbar to create a smooth, convex segmentation around your region of interest. Use the Model Marker Placement Tool to lay down border points. Once you place more than three border points, a best-guess segmentation will automatically be created from your border points. You can add as many border points as needed, with the segmentation guess automatically adjusting as necessary. You can move existing border points by de-selecting the Model Marker Placement Tool, and clicking and dragging using the basic mouse tool. The model will update accordingly.
-*'''Segmentation of the fixed image'''
+*'''Model-Based ROI''': Use this drop-down menu to switch between different sets of markup points. Additional points added with the ROI toolbar will update whichever model is selected in this drop-down menu. Pre-existing sets of points can also be loaded into this menu and turned into the template for a segmentation.
-*'''Moving image''' (optional): image to be registered
+<gallery>
-*'''Segmentation of the moving image'''
+D_Model_Segmentation_Step_4A.PNG|Step 4 User Interface
-*'''Registration affine transform''': transform to store the result of affine registration
+D_Model_Segmentation_Step_4B.PNG|Step 4 Example ROI from Markups
-*'''Registration deformable transform''': transform to store the result of deformable registration
+</gallery>
-*'''Visualization''': initializes slice viewers to show overlay of the fixed and moving images (before or after registration), outline of the fixed image segmentation and the registration transform grid and visualization of the fixed and moving segmentation label surfaces in 3d view
+== Step 5. Thresholding ==
+*'''Threshold Range''': Adjust this slider to set a simple intensity threshold on the smooth segmentation created in the previous step. Voxels in the newly-created label named "[VolumeName1]_roi_label_thresholded" with intensities outside of the slider's bounds will be erased. The original segmentation without any thresholding is available under a second label named "[VolumeName1]_roi_label", and a version of the original volume cropped to that label is available under the name "[VolumeName1]_roi_cropped".
 <!--
 {{documentation/{{documentation/version}}/module-parametersdescription}}
@@ Line 74: / Line 99: @@
 <!-- ---------------------------- -->
 {{documentation/{{documentation/version}}/module-section|Similar Modules}}
-Volume Clip
+Volume Clip - https://www.slicer.org/wiki/Documentation/Nightly/Extensions/VolumeClip <br>
-ChangeTracker
+ChangeTracker - https://www.slicer.org/wiki/Documentation/Nightly/Extensions/ChangeTracker
 <!-- ---------------------------- -->
 {{documentation/{{documentation/version}}/module-section|References}}
 <!-- ---------------------------- -->
 {{documentation/{{documentation/version}}/module-section|Information for Developers}}
-* Source code: https://github.com/QTIM-Lab/3D_Model_Segmentation
+* Source Code: https://github.com/QTIM-Lab/3D_Model_Segmentation
 <!-- ---------------------------- -->
 {{documentation/{{documentation/version}}/module-footer}}
 <!-- ---------------------------- -->

@@ Line 56: / Line 56: @@
 *'''Registration Method''': Choose which registration method you would like to use. "No Registration" allows you to skip this step. "Rigid Registration" will attempt to rotate and translate the moving image into the space of the fixed image. "Affine Registration" will rotate, translate, scale, and skew the input image. Both affine and rigid registration methods will produce a 3D Slicer Linear Transformation object which you can inspect in the "Transforms" module. "Deformable Registration" will perform deformable BSpline registration on the moving image, and correspondingly create a BSpline Transform object. In general, each method will be more accurate but more time-consuming than the previous method.
 *'''Registration Output''': Either create a new volume for the post-registration moving image, or replace the original image. The new volume will have the name "[VolumeName1]_reg_[VolumeName2]".
 *'''Segmentation of the fixed image'''